The New Market Makers: Who Builds, Prices and Backs Onchain Finance
Updated: 12 hours ago
How HIP-3, Trade.xyz, Hyperliquid, Ondo and Aster are rebuilding the machinery of trading—and why the decisive differences lie beneath the price chart.
Research and analysis · September 11, 2026
Based on public platform documentation and primary sources. Hypothetical examples are identified as such.
A geopolitical shock breaks on a Saturday. Oil traders have an immediate reason to reconsider prices, but the benchmark futures exchanges are between sessions. On an onchain derivatives venue, orders can continue to arrive. Buyers raise their bids. Sellers withdraw offers. A price moves before the conventional market reopens.
That price is economically meaningful: somebody is prepared to risk capital at it. But its meaning requires care. It could represent expectations for the next futures session, demand for weekend insurance, a shortage of sellers, or the mechanical consequences of leveraged positions being liquidated. Several of those forces can operate simultaneously.
The emerging contest over continuous markets therefore reaches beyond opening hours. It concerns who can create a financial contract, how its reference price is constructed, what supports its value, and how the system handles losses when collateral proves insufficient.
Hyperliquid’s HIP-3 framework, the markets deployed by Trade.xyz, Ondo’s tokenized securities platform and Aster’s derivatives infrastructure approach these questions differently. They can display similar-looking exposure to oil, equities or other assets while creating materially different obligations for their users.
The most useful distinction is between a market for price exposure and a token supported by an underlying investment.
A perpetual derivative records a leveraged economic position. A backed security token connects an onchain instrument to assets and legal arrangements outside the blockchain. Their charts may converge, but the infrastructure enforcing that convergence is different.
HIP-3—Hyperliquid Improvement Proposal 3—allows qualifying builders to deploy perpetual markets on Hyperliquid. The builder defines the contract and oracle and operates parameters such as leverage and potential settlement. HyperCore supplies the trading and margin infrastructure. The published specification requires 500,000 HYPE staked by a deployer, although that requirement may change. Each deployed perpetual exchange has independent books, margining and settings.
Trade.xyz, branded trade[XYZ], occupies the product-building layer. It deploys HIP-3 perpetuals and provides an interface for interacting with Hyperliquid markets. Its documentation expressly distinguishes that interface from HyperCore: the website does not operate HyperCore or hold user assets, and it is not the exclusive route to Trade.xyz’s markets. These are separate responsibilities within a shared trading system.
This architecture changes the economics of launching an exchange product. A builder can focus on choosing exposures, sourcing prices, specifying contracts and attracting liquidity while using an existing execution environment. Hyperliquid, in turn, can host a wider range of markets without originating every contract itself.
It also makes product-level diligence essential. Shared infrastructure does not establish that every market has equally resilient pricing, equally deep liquidity or equally conservative risk settings. Two contracts on the same blockchain can have very different failure modes.
HIP-3’s staking requirement adds an economic commitment, but it should not be mistaken for customer insurance. The protocol describes slashing principally in terms of harmful technical behavior and protocol integrity. Slashed stake is burned, rather than distributed to affected traders. The deployer can also use a protocol action to halt trading and settle positions at the current mark.
Underneath the interface, HyperCore maintains an order book for each asset. Orders match by price-time priority. Its clearinghouse tracks balances and positions, while margin checks occur when orders are opened and again for resting orders when they match. This coordination matters because a resting order can become unsafe after prices move, even though it was adequately collateralized when submitted.
The word “clearinghouse” needs context here. HyperCore uses it for a component of its execution state. In traditional futures, a central counterparty is also a legal institution embedded in a network of clearing members and enforceable obligations. Similar vocabulary does not establish identical protections. ICE’s Brent specification, for example, describes ICE Clear Europe’s counterparty role and its contractual relationship with clearing members.
Pricing is another set of distinct functions concealed by a single chart.
A feed distributes observations or calculated values. An oracle makes selected outside information available to the trading system under a defined methodology. The risk engine then decides how that information affects margin, funding and liquidation. In this context, “oracle” refers to data infrastructure, not Oracle Corporation.
Four prices deserve separate names. The underlying-market price is an observation from the referenced asset or contract. The oracle or index price is the value selected or constructed for the derivative. The last-traded price is the price of the venue’s most recent execution. The mark price is a valuation input used by the risk system. Depending on the product, these values can differ substantially.
A reliable oracle therefore needs more than a number. It needs an identified instrument, currency, timestamp and session; rules for aggregating contributors; and a response to missing, delayed or implausible observations. A Brent futures quote for one delivery month is not interchangeable with another month, a physical assessment or a WTI quote simply because all are denominated in dollars per barrel.
Pyth provides a useful illustration. Its Pro documentation describes aggregating publisher prices by median, subject to a minimum contributor count. If that minimum is not met, the feed can carry forward its previous aggregate. Consumers must examine the feed-update timestamp to distinguish fresh pricing from carried-forward data. Its confidence metric describes disagreement among publishers; it is not a promise about future volatility or a guaranteed execution range.
That distinction becomes critical during closures. A network can remain operational while receiving no new observations from the primary market. Transporting an old price quickly does not make the observation current. Equally, a fresh estimate from a continuously trading alternative venue can be useful without being the official price of a closed exchange.
Pyth identifies Trade[XYZ] among the HIP-3 venues using its pricing infrastructure. This establishes a data-provider relationship, not a universal contract formula: HIP-3 delegates oracle design to market builders, and different feeds or instruments can require different treatment.
Trade.xyz’s published discovery-bound mechanism demonstrates how continuous trading can coexist with restrictions on valuation. It limits the mark within a band around a reference price, with instantaneous width expressed as plus or minus one divided by maximum leverage. Certain markets permit the reference to re-anchor when the oracle approaches a trigger, shifting the band in steps. Once the permitted resets are exhausted, a hard boundary applies until external pricing resumes. The documentation’s illustrative WTI case uses a 5% band; that example should not be treated as every oil market’s current setting.
The economic trade-off is straightforward. Bounds constrain how far a risk valuation can move under reduced external information. They also constrain what that valuation can express. A price pinned to a boundary may reveal a limit in the market’s rulebook rather than a consensus that the asset is worth exactly that amount. Re-anchoring expands the possible discovery range but does not remove the eventual reopening risk.
Funding provides a separate connection between perpetuals and their reference. Without scheduled expiry, a perpetual needs an incentive for traders to close a persistent premium or discount. Periodic transfers between long and short positions help supply that incentive. Hyperliquid’s documentation describes an interest component and a premium component, with funding paid hourly. The premium calculation uses impact prices derived from order-book depth.
Funding encourages convergence; it does not guarantee immediate convergence. A trader considering the opposite side of an expensive perpetual must weigh prospective funding receipts against the possibility that the price keeps moving, the hedge is unavailable, or collateral is exhausted before prices converge.
Consider a purely hypothetical $10,000 long position paying 0.01% every hour. Holding notional and rate constant, its funding cost would be $1 an hour, $24 a day and $720 over 30 days. Actual rates and position values change, but the example explains why a correct directional view can still produce a disappointing return. Perpetual exposure has a financing path as well as a price path.
Oil makes the consequences of contract design especially visible.
Brent and West Texas Intermediate are related crude benchmarks with different physical and commercial foundations. Their relative prices respond to quality, location, transportation and supply conditions. The Brent–WTI spread is itself a traded economic relationship, rather than a discrepancy that must disappear.
CL is the conventional NYMEX WTI futures code. A conventional CL position is tied to a dated futures contract and the exchange’s settlement obligations. CME describes its benchmark WTI contract as physically settled. ICE Brent futures have their own terms: a 1,000-barrel contract, defined expiration, and delivery based on exchange for physical, with an option to cash settle against the ICE Brent Index. Describing every oil future as simply “cash-settled oil” loses material distinctions.
A perpetual displayed as BrentOil, BRENTOIL, CL or another WTI label must instead be understood through its venue-specific specification. A familiar ticker does not give its holder an ICE or NYMEX position, access to that exchange’s clearing arrangements, or a right to receive crude. The economic exposure depends on which benchmark the oracle follows and how the perpetual handles financing and settlement.
This creates several layers of basis risk—the possibility that related instruments move differently. Brent versus WTI is benchmark basis. One delivery month versus another is calendar basis. A perpetual versus its reference is venue or instrument basis. A weekend estimate versus the next primary-market opening adds another source of divergence.
Those differences explain why two “oil” charts can legitimately disagree. Before attributing a gap to manipulation, the comparison must align the benchmark, contract month, currency, timestamp, session and price type. Comparing Friday’s settled future with Saturday’s last perpetual trade answers a different question from comparing simultaneous executable quotes.
There is also a common misconception about rolls. A perpetual has no routine expiry for its holder. That fact alone does not eliminate the economics of a rolling reference. If an oracle follows dated futures and changes its reference month, the transition methodology still matters. Trade.xyz’s documentation index includes reference-contract roll schedules, underscoring that expiry and reference construction are separate subjects.
Suppose, hypothetically, the near future is $75 and the following month is $77. The curve is in contango. Reversing those prices produces backwardation. A switch between contracts, a weighted blend, a price adjustment and an explicit funding adjustment can distribute that difference differently. None can make the economics irrelevant merely by keeping the same ticker on screen. This is a general explanation, not a claim about an unverified BrentOil roll formula.
Continuous oil pricing is becoming a product in its own right. Pyth announced a 24/7 Oil Index in March 2026, describing a composite assembled from institutional and onchain sources. It is an example of how price infrastructure can incorporate activity that persists after traditional venues close. Its existence, however, does not establish that every Trade.xyz oil contract uses that particular index.
Such composites raise a further design question: how independent is the reference from the market consuming it? If a derivative contributes substantially to an index that then influences its own mark or funding, feedback can develop. That is a structural issue to investigate through constituent weights and methodology, not evidence that a specific platform is manipulating prices.
The strongest test of weekend price discovery is consequently empirical. Does the venue support meaningful trade size? Does its price anticipate subsequent benchmark trading? Does it remain informative after accounting for funding, liquidation flow and thin liquidity? A visible print is evidence of a transaction. A dependable benchmark requires more.
Ondo approaches market access through a different balance-sheet structure.
For this comparison, the relevant business is Ondo Stocks, offered through Ondo Global Markets, rather than the ONDO governance token. Its stock tokens are designed to provide economic exposure to underlying securities and reinvested distributions, net of applicable withholding taxes. They do not give holders direct title to the referenced shares or ordinary shareholder voting rights.
The distinction from a perpetual is fundamental. Holding an unleveraged stock token does not inherently require periodic long-to-short funding or expose the holder to a perpetual exchange’s maintenance-margin liquidation. If that token is later borrowed against or used inside a leveraged protocol, however, the additional arrangement can introduce both financing and liquidation risk.
Ondo describes an issuer structure separated from its technology provider, underlying securities held at registered custodial broker-dealers, additional collateral, and a third-party security agent with rights over backing assets. Its documentation names Ankura Trust Company as verification and security agent. These arrangements seek to connect tokenholder claims to assets and a recovery process. They do not make custody, issuer operations or legal enforcement disappear.
The execution model also differs. Ondo supplies transaction quotes informed by inventory and market conditions. Its developer interface can return a signed, time-limited attestation authorizing a mint or redemption. This connects an offchain pricing decision to an onchain transaction. An attestation authorizing a trade should not be confused with an independent attestation that verifies backing assets.
Its settlement plumbing illustrates the role of inventory. Ondo documents USDon as the platform’s purchase asset, with a swapper that can convert USDC and USDon within a transaction. Immediate conversion depends on available swapper liquidity. Atomic exchange means the onchain transaction’s component steps succeed together; it does not mean the external brokerage, banking and securities systems have become instantaneous or unlimited.
Token valuation introduces another easily missed distinction. Because these instruments track total returns, one token can come to represent exposure to more than one underlying share. Corporate-action adjustments and display conventions therefore matter when comparing token and stock prices. A token priced above the share price is not automatically trading at a speculative premium.
Ondo’s documentation now describes predominantly 24/5 availability with selected assets enabled for additional off-hours trading. During those periods, proprietary quotes can carry wider spreads and conservative size limits. Blockchain transferability, secondary-market trading and direct mint/redemption availability are separate capabilities; an investor should not infer one from another.
Its competitive proposition is therefore partly institutional: connect onchain distribution to backing, servicing and redemption. A perpetual venue can create exposure without acquiring one matching share for every unit of notional. A backed-token issuer must maintain the assets and operational arrangements supporting its obligations. These models can serve overlapping users while carrying different costs and constraints.
Aster is the closer derivatives competitor, but its product family requires careful separation.
The platform documents order-book perpetuals alongside Shield Mode and its separate 1001x offering. Their liquidity and execution designs are not interchangeable. A claim about one mode should not be generalized to every product carrying the Aster name.
Aster’s stock-perpetual documentation describes USDT settlement and index inputs from multiple providers, including Pyth. Its standard mark is the median of three values: a funding-adjusted index, an index adjusted by a five-minute average basis, and the contract price. During designated low-liquidity periods, it replaces the raw contract-price input with an exponentially weighted average.
The aim is to reduce the influence of isolated trades on valuation.
Using a median reduces the influence of one extreme input; smoothing reduces sensitivity to brief spikes. Neither manufactures external liquidity. Both also introduce trade-offs: a deliberately stable mark can adjust more slowly when a large move is economically justified. Traders therefore need to distinguish the price at which they can execute from the value used to assess their solvency.
Aster also handles equity dividends through a specific derivative mechanism. Its published rules describe a special funding settlement in which shorts compensate longs, alongside temporary trading restrictions and changes to price-deviation controls. That contrasts with Ondo’s total-return token accounting. Neither mechanism turns a perpetual trader or tokenholder into an ordinary registered shareholder.
Collateral flexibility is another competitive dimension. Aster’s multi-asset mode recognizes eligible collateral at specified value ratios, nets positions through cross margin, and includes automatic conversion rules. Discounting collateral acknowledges that its market value may not be fully available during stress. Yield-bearing or volatile collateral can improve capital utilization while adding dependencies to the trading account.
A hypothetical account holding volatile crypto collateral while long equities illustrates the risk. A broad sell-off can reduce both the collateral’s value and the derivative’s equity at once. Capital efficiency in ordinary conditions can become accelerated margin pressure in a correlated shock.
Aster’s newer chain documentation describes a purpose-built Layer 1 using proof-of-staked authority and privacy features involving encrypted orders, zero-knowledge verification and stealth addresses. It also states that core chain and RPC components are not currently open-source and that external validator participation is unavailable during Phase 1. These are material qualifications to evaluate alongside its performance and decentralization claims. The architecture page describes privacy as an account setting, so it should not be portrayed as uniform across all activity.
Privacy can reduce information revealed about a trader’s strategy. Public visibility can help outsiders inspect market behavior. The practical comparison is about which information is exposed, who can verify hidden activity, and which operators or validators retain control. A “DEX” label by itself does not answer those questions.
Aster has also documented AOS-2, a perpetual-listing route requiring eligibility checks, a designated market maker, a one-million-ASTER stake locked for four years, a validator vote and risk-team configuration. Although described as permissionless, that documented process includes approval stages. It differs from HIP-3’s model of builder-operated markets on shared infrastructure and should not be described as an identical listing system.
The infrastructure differences can be condensed as follows. The comparison concerns the products discussed here, rather than every service offered by each organization.
Dimension | Trade.xyz / HIP-3 | Ondo stock tokens | Aster perpetuals | Conventional oil futures |
Position held | Builder-defined perpetual exposure | Backed token tracking security economics | Perpetual derivative exposure | Dated exchange contract |
Execution | HyperCore order books | Issuer quotes and secondary venues | Product-dependent; order books for the perps discussed | Exchange markets and permitted trade facilities |
Value anchor | Deployer oracle, funding and market rules | Backing, quotes and redemption | Index, mark and funding rules | Contractual settlement or delivery |
Principal dependencies | Builder, data, chain and liquidity | Issuer, custodian, servicing and contracts | Data, operators, chain and liquidity | Broker, clearing member, exchange and CCP |
Holding economics | Price change, funding, fees and reference design | Total return, spreads and applicable deductions | Price change, funding, fees and adjustments | Futures price, financing and any roll |
Comparison synthesizes the linked specifications above. Contract-level rules prevail over platform-level descriptions.
The most consequential differences appear when a position fails. Hyperliquid describes liquidation when equity falls below maintenance margin, with attempts to close positions through its book. Its auto-deleveraging rules provide a final mechanism that can close opposing positions when an account becomes negative.
Aster likewise documents ADL as a response when other protective mechanisms cannot cover losses in time. A profitable trader’s position can consequently be reduced because another part of the system has failed.
This is where price quality, collateral quality and liquidity become inseparable. A defensible mark is insufficient if a large position cannot be closed near it. Deep liquidity is insufficient if the reference is faulty. Ample collateral in one asset can become inadequate if that asset depegs or cannot be converted when needed.
Traditional markets face related problems, but their institutional response is organized differently. Exchange rules, clearing-member obligations and contractual settlement sit alongside the trading technology. Onchain systems can automate parts of that machinery and make some activity easier to inspect. The underlying questions—who owes what, what resources support the obligation, and what happens after a default—remain.
Nor is the competition simply continuous crypto markets against traditional exchanges open only during office hours. Benchmark oil futures already trade across extensive electronic sessions. CME’s WTI page also advertises a proposed 10-barrel product with 24/7 access, explicitly subject to regulatory review. That is an announced direction, not proof that the incumbent CL contract has become continuously available.
As opening hours converge, other advantages become more decisive: executable liquidity during stress, reliable feeds, transparent product definitions, predictable collateral treatment and access to capital that can absorb temporary imbalances. Cheap transactions and fast blocks help a market function, but they cannot substitute for those foundations.
HIP-3’s opportunity lies in enabling specialist builders to create markets on shared trading infrastructure. Trade.xyz’s opportunity lies in making those contracts useful, liquid and intelligible. Ondo’s opportunity lies in connecting securities backing and servicing to onchain distribution. Aster’s opportunity lies in combining derivatives execution, collateral flexibility and privacy features into a competitive trading environment.
The Saturday oil price is where those designs meet a demanding test. Somebody must supply the information. Somebody must quote the other side. Somebody must finance the position. And the rulebook must specify what happens if the move exceeds what the system can safely absorb.
A market earns its importance when people can rely on that machinery. Keeping the screen illuminated through the weekend is the beginning.
Research note.
Sources were reviewed on September 11, 2026. Platform descriptions are attributed to their published documentation; this article does not independently audit reserves, code, validator operation, execution quality or performance claims.
Trade.xyz’s documentation was being reorganized during research: its overview, directory and discovery-bound material were retrievable, but several detailed oracle and contract-specification pages were not. Accordingly, exact current BrentOil/WTI feed identifiers, reference months, roll formulas, leverage settings and reset counts are not asserted here.
The oil examples explain contractual differences and general economics. Access restrictions and governing product documents remain applicable; technical availability does not establish eligibility in a particular jurisdiction.
Comments